We move Brazil USD-denominated sovereign bonds to Neutral from Underweight following the first round and remain defensive on the curve. Our September Underweight combined valuation and event timing, and we identified the first round as the point to reassess it. Flávio Bolsonaro finished ahead of President Lula with 47.0% of valid votes against 45.2%. Roughly 8% went to eliminated candidates, mostly from the center and right of center. Right-of-center parties also strengthened their position in Congress, and the Liberal Party (PL) emerged as the largest party in both chambers, with 28 senators and 121 deputies. That gives a potential Bolsonaro administration a stronger legislative starting point for fiscal reform.

Neither of the re-entry conditions we set in September has been met. The Brazil 2035s trade at 199 bps of OAS, not around 230 bps, and no fiscal framework has been delivered. Those thresholds were set before the first-round result. The first round removes most of the event-risk rationale for the Underweight, but valuation remains restrictive. The Brazil 2035s at 199 bps offer only an 8 bp pickup to the Mexico 2034s at 191 bps, with 6.3 versus 5.8 years of duration. At the front end, the Brazil 2029s trade at 51 bps against 62 bps on the Mexico 2029s, with reported durations of 2.4 and 2.3 years. Further out, the Brazil 2041s trade 34 bps through the Mexico 2040s. The political signal is better; the valuation is worse. That leaves too little upside for Overweight and too little evidence for a broad sovereign Underweight.

Since September 17, Brazil has tightened 9 to 16 bps across the belly and long end while peers widened, with the Mexico 2034s 16 bps wider and the South Africa 2036s 15 bps wider. The clearest relative-value dislocation is the Brazil 2041/Mexico 2040 pair. The Brazil 2041s moved from 6 bps through the Mexico 2040s on September 17 to 34 bps through today, an additional 28 bps of relative tightening in three weeks, with reported durations of 9.2 and 8.6 years, respectively. The Brazil 2035/Mexico 2034 pair shows the same 28 bp compression, from a 36 bp Brazil premium on September 17 to 8 bps today, with durations of 6.3 and 5.8 years. Both Mexico bonds widened over the period, so two actual-bond pairs showing the same 28 bp move provide clear evidence of Brazil-specific repricing.

Read more

We maintain our Opportunistic Overweight recommendation on CSN, with lower conviction, and shift our preferred bond to the CSN Inova 11.000% 2030s from the CSN Resources 8.875% 2030s. The guidance cut does not create CSN’s liquidity problem. It weakens the principal operating buffer that was giving the company time to solve it. The reduction in 2026 production guidance and increase in C1 cash-cost guidance at CSN Mineração (CMIN) are credit negative. They make the thesis more dependent on the cement sale, working-capital release and continued refinancing access. The curve, however, has repriced by more than our fundamental view has changed. The Resources 2030s have widened 310 bps since our August report to 1,729 bps OAS. The Inova 2030s now trade at $71.3, 1,910 bps OAS and a 22.7% YTW. On a simple coupon-carry basis, they break even over 12 months at roughly $60.3. Default remains outside our base case over a 9 to 12 month horizon, and we think current levels compensate for the higher execution risk.

The bond switch reflects a material change in relative value since August. The Inova 2030 notes now offer 181 bps of additional OAS and 160 bps of additional YTW over the Resources 2030s, with 0.2 years less duration, and their dollar-price premium has narrowed to 3.1 points from 6.3 points when we last reviewed the curve on August 19. For investors assigning greater weight to an impairment scenario, the long end remains the defensive expression, and we prefer the CSN Resources 5.875% 2032s to the 4.625% 2031s. We continue to avoid the residual CSN Inova 6.750% 2028s.

Read more

We initiate coverage of the Province of San Juan with an Overweight recommendation on the USD600mn 9.55% senior unsecured notes due October 2035, with an entry level of at least 50 bps over Córdoba 2035. The bond priced on September 24 at $98.234 to yield 9.875%, about 54 bps over Córdoba 2035 on September 28 levels. Córdoba 2035 has since widened to 9.482% YTW and 481 bps OAS from 9.338% and 466 bps, so our 50 bps entry now corresponds to a SANJUA yield near 9.98% and a price near $97.66. We do not treat the 39 bps pickup implied by the frozen reoffer yield as a current valuation. If the bond opens with full pass-through of Córdoba’s move, near 10.02% and $97.47, the original concession would be restored. If it opens near reoffer, we would hold existing positions but not add.

San Juan is not a top-tier provincial credit on current metrics. Our proprietary provincial framework ranks it 9th of 15 provinces, reflecting the post-issuance increase in leverage and limited provincial tax revenue. The Overweight rests on credit resilience. We think the province can carry the larger debt burden through the mining investment cycle without relying on the copper projects to generate material new revenues before 2033.

San Juan’s structural ratios deteriorate materially after the issue. Pro-forma debt reaches 9.7% of provincial gross geographic product (PBG) and 46.5% of revenue, while provincial tax revenue is only 12.6% of total revenue. Those metrics compare poorly with Córdoba, Santa Fe and Neuquén and justify a clear spread premium. At the same time, the new bond adds USD57.3mn of annual coupon expense through 2032, recent fiscal results remain positive and the province entered the transaction with substantial liquidity. We estimate total 2027 debt service, including existing local and multilateral obligations plus the new SANJUA coupon, at approximately USD86mn. The USD168.4mn of liquidity disclosed in May 2026 is close to 2.0x that amount before considering fiscal resources generated during the year. Most of that liquidity is peso-denominated, so the strength of the buffer still depends on national FX policy and the province’s ability to convert local-currency resources into dollars. That exposure makes the 2027 presidential election cycle directly relevant to the credit even if provincial fiscal performance remains strong.

Debt-service capacity remains credible under a deliberately conservative stress. If San Juan’s 2025 current fiscal balance of Ps341.2bn were held flat in nominal pesos through 2032 while the peso followed the depreciation assumptions used in the prospectus debt schedule, the resulting amount would still equal roughly 2.0x estimated annual pro-forma debt service in 2027 and about 1.5x in 2032. This is not a forecast. It is an EM Spreads stress test designed to assess whether the province can service debt through the capex period without assuming revenue growth from the new mining pipeline. The 1H26 current balance reached Ps212.5bn, providing further evidence that recurring fiscal capacity remained intact at the time of issuance.

Read more

We move our recommendation on Braskem to Underweight from Neutral and shift our preferred expression across the senior unsecured curve to the BRKME 5.875% 2050s from the BRKME 8.500% 2031s. The downgrade is primarily a valuation call. Creditors are reportedly resisting Braskem’s latest proposal, talks have become more contentious again after the parties agreed to pursue the extrajudicial restructuring route, and recuperação judicial has returned more prominently to the discussions. At the same time, the reported terms give bondholders the first concrete indication of how Braskem may attempt to reduce debt rather than only extend maturities and capitalize interest. Yet the 2028-2041 senior curve is down only about 2.5 points on a face-weighted basis since our August 2Q26 report despite the more creditor-punitive restructuring framework, while the 2050s have fallen 5.6 points. The 2028s through 2041s currently trade between $51.4 and $58.2, with a face-weighted average of approximately $54.7, versus $43.3 for the 2050s. We rate the 2028s through 2041s Underweight and the 2050s Neutral. We prefer the 2050s because the lower dollar price provides more protection if restructuring treatment converges, while recognizing that the advantage can reverse if bond-specific coupons and maturities are preserved.

Braskem’s latest proposal reportedly asks creditors to provide approximately US$2.0bn of new money. About US$1.25bn would fund a repurchase of existing debt at up to 50% of face value, while US$750mn would fund working capital. The proposal at least addresses principal reduction, unlike Braskem’s original framework. The economics, however, still appear insufficient to create a durable balance-sheet solution if the new money is itself debt, reinforcing our view that Braskem ultimately requires both substantive creditor concessions and fresh shareholder capital. A group of bondholders is now reportedly asking Petrobras and IG4 to contribute approximately US$3.0bn of fresh equity, while multiple bondholders have indicated that they do not intend to provide additional funding themselves. We see the demand as further evidence that the current proposal remains far from consensual. It does not mean US$3.0bn of shareholder capital is likely to be provided.

The downgrade does not depend on the precise reported proposal terms. In August, the absence of a large collateralized layer ahead of the bonds was one of the main reasons we remained Neutral. That protection is now less secure. Negotiations contemplate creditor-funded new money, while proposals under consideration have included security interests over assets. Prices across most of the curve still provide limited recovery cushion, while the 2050s offer more price protection but remain exposed to the same unresolved restructuring risks.

Read more

We move Brazil USD-denominated sovereign bonds to Underweight from Neutral. This is a valuation and event-timing call into the October 4 first round and the October 25 runoff, not a change in our credit assessment. Brazil’s external position is stronger than at our July initiation. Gross international reserves are near $370bn and projected to increase through 2028, FDI covers the current-account deficit and the sovereign funds predominantly in local currency. What has changed is the price, while the medium-term growth and rates mix has become less supportive of debt stabilization. Since initiation the 2029s have compressed 37 bps to 29 bps of OAS, the 2030s 32 bps to 94 bps, the 2031s 25 bps to 119 bps and the 2035s 17 bps to 209 bps. Over the same period, the LatAm Broad Bond Index USD compressed 8 bps to 231 bps. At current levels, Colombia trades wider than Brazil across most comparable points, while Mexico remains tighter. Every major point on the curve still trades within 8 to 19 bps of its one-year OAS minimum. The compression accelerated in September as several polls narrowed toward statistical ties and prediction markets moved toward the challenger. Over the same period our economic monitor’s 2027 growth forecast moved from 1.7% to 1.3% and its end-2027 Selic forecast from 10.6% to 12.3%, while the deficit, interest bill and debt projections remained unchanged.

We read that as the market pricing a change in fiscal regime that has not been delivered, by an administration that has not been elected, through a Congress that has not been seated. Prediction markets have moved further toward the challenger than the polling has. We treat that divergence as a positioning and event-risk signal, not as a forecast of the result. For the bonds, the relevant question is how much fiscal improvement current spreads already discount, not which candidate wins. Spreads at these levels provide some peer cushion but limited protection against adverse post-election fiscal outcomes. Two of the triggers we set at initiation have fired. We said we would reduce the belly preference if the 2035s and 2036s compressed without better fiscal evidence, and turn more defensive on fiscal slippage or a shift in the easing path. The belly compressed 8 to 17 bps without a corresponding improvement in the fiscal outlook. Fuel-relief measures carried a roughly R$40bn fiscal cost between March and September, and on September 17 the government announced a 15% increase in Bolsa Família benefits with an estimated additional cost of R$22bn in 2027. Investment and government consumption supported second-quarter growth while private consumption contracted, and the central bank is easing into 2027 inflation expectations that have moved higher. Gross debt still reaches 100% of GDP in 2027 on the general government definition in our economic monitor.

The Underweight is not a forecast of the election result. At current levels, valuation cushion is meaningful and the ordinary-case payoff is broadly balanced. Credible fiscal consolidation after the election could take the 2035s toward the Mexico 2034s at 176 bps, roughly 33 bps of compression and about 2.1 points on 6.4 years of duration. Fiscal slippage could instead retrace them toward the three-month average of 219 bps or the one-year average of 237 bps, a loss of roughly 0.6 to 1.8 points, with the one-year maximum at 278 bps implying about 4.4 points in the tail. The Underweight carries moderate conviction because the 2035s remain 21 bps inside our 230-bp Neutral re-entry area, sit only 18 bps above their one-year OAS minimum and 28 bps inside their one-year average, and the fiscal improvement required to sustain further country-risk compression has not been delivered. A constructive post-election fiscal signal could compress spreads quickly, but sustaining that move would require execution. Fiscal disappointment can reprice the curve before delivery becomes visible in the data.

Read more

We maintain our Overweight recommendation on Pemex, with lower conviction, and change our preferred expression for new money to the 7.690% 2050s. We retain the 6.625% 2035s as the core lower-duration position for accounts that do not want long-end volatility. For accounts with duration flexibility, we would rotate front-end positions into the 2035s or 2050s; we would retain the front end only for short-duration mandates prioritizing downside protection. We view Mexico’s proposed 2027 budget as neutral to modestly positive for the credit. The P$81.1 billion transfer allocated to Pemex debt service is approximately 69% below the P$263.5 billion budgeted for 2026, but we do not interpret the reduction as a withdrawal of sovereign support. It primarily reflects the decline in near-term refinancing needs produced by Mexico’s liability-management strategy rather than an improvement in Pemex’s standalone cash generation. Explicit support remains in the budget, while the 2027 financing framework preserves substantial domestic and external borrowing authorization.

The more consequential change for bondholders since our 2Q26 report is where the refinancing improvement is now reflected in the curve. The front end has tightened sharply, while intermediate and long-dated bonds have widened. The 5.350% 2028s are 34 bps tighter at 75 bps OAS and the 8.750% June 2029s 22 bps tighter at 96 bps, while the 6.625% 2035s are 12 bps wider at 315 bps and the 7.690% 2050s 21 bps wider at 433 bps. The curve is now distinguishing more clearly between the refinancing risk sovereign action has reduced and the longer-term risks that remain, including dependence on support, weak standalone cash generation and operating execution.

Read more

We maintain our Neutral recommendation on MercadoLibre (Baa3/BBB-/BBB-), unchanged from our 2Q26 report, and continue to prefer the MELI 3.125% 2031s. The new 5.850% 2036s improve liquidity and extend the debt profile, but at a 6.139% reoffer yield we do not think the bond offers enough compensation for approximately 7.5 years of duration and a funding-intensive balance sheet. The credit itself remains strong. MercadoLibre continues to deliver rapid operating growth, has demonstrated access to long-term funding and now has better coverage of short-term obligations. The constraint remains valuation relative to cash generation, leverage and the funding requirements of Fintech. The credit portfolio grew 75.2% YoY in 2Q26 against 49.8% consolidated revenue growth, while total debt increased 46.7%. Adjusted FCF was only US$214 million despite US$3.22 billion of free operating cash flow, while gross and net leverage stood at 3.39x and 1.65x. We therefore view the new 2036s as a Hold. The 2031s remain our preferred expression given their $90.3 price and 3.9-year duration, while for accounts seeking longer duration we prefer the new benchmark to the 2033s.

MercadoLibre priced US$1.0 billion of 5.850% senior unsecured notes due September 14, 2036 at $97.864 to yield 6.139%, with an issue spread of 130 bps. Pricing tightened approximately 30 bps from initial price talk around 160 bps, and the company said the transaction attracted more than 100 institutional investors. Settlement is scheduled for September 14. The notes are SEC registered, governed by New York law and guaranteed by six operating subsidiaries across Argentina, Brazil, Chile, Mexico and Colombia. They are callable at a T+20 make-whole spread before June 14, 2036 and at par thereafter. Proceeds are for general corporate purposes.

New Issue Terms

• Issuer: MercadoLibre, Inc.

• Instrument: Senior unsecured notes

• Size: US$1.0 billion

• Maturity: September 14, 2036

• Coupon: 5.850%, semiannual

• Issue Price: $97.864

• Reoffer Yield: 6.139%

• Issue Spread: 130 bps, from IPT around 160 bps

• Duration: approximately 7.5 years, EM Spreads estimate

• Expected Ratings: Baa3 / BBB- / BBB-

• Settlement: September 14, 2026

• Optional Redemption: Make-whole at T+20 before June 14, 2036; par thereafter

• Use of Proceeds: General corporate purposes; proceeds are expected to further strengthen liquidity

• Format: SEC Registered

• Governing Law: New York

• Expected Listing: Nasdaq Bond Exchange

• ISIN: US58733RAJ14

• Guarantors: MercadoLibre S.R.L., Mercado Livre Brasil Ltda., MercadoLibre Chile Ltda., DeRemate.com de México, MP Agregador and MercadoLibre Colombia Ltda.

Read more

We maintain our Neutral recommendation on YPF (B1/B/B-) and shift our preferred expression to the new 7.550% June 2035 senior unsecured notes from the YPFDAR 9.500% 2031s. The increase in the tender cap makes the transaction more credit positive than at launch, as YPF can now fully retire the targeted 2027 bonds and meaningfully reduce the 2029s ahead of the presidential election in October 2027. YPF’s underlying credit trajectory remains positive. Valuation has also improved since our 2Q26 report. We continue to see the opportunity as concentrated in bond selection rather than broad enough across the curve to support a renewed issuer-level Overweight. At the bond level, we prefer the new 2035s, which now trade approximately 29 bps above YPF’s curve on a YTW basis. The 9.500% 2031s, meanwhile, no longer offer the spread pickup that made them our preferred instrument in our 2Q26 report. We retain the 8.250% 2034s as our preferred shorter-duration alternative and continue to avoid the 7.000% 2033s and 7.000% 2047s.

YPF priced US$1.2 billion of 7.550% June 2035 senior unsecured notes at a 7.85% yield and $98.075 issue price, with settlement expected September 18, 2026. On pricing day, the company doubled the Maximum Purchase Price for its concurrent tender offers to US$1.0 billion from US$500 million, excluding accrued interest, while stating that no other tender terms had changed. The 6.950% 2027s therefore remain Acceptance Priority Level 1 at $101.75 and the 2.500%/9.000% Step Up Amortizing 2029s Priority Level 2 at $104.20, each per US$1,000 principal amount. If fully utilized and supported by sufficient participation, the enlarged cap could retire all US$643 million of the 2027s and approximately US$331 million of the 2029s, for combined principal retirement of roughly US$975 million. The tender expires September 16 and is expected to settle on or around September 18.

New Issue Terms

· Issuer: YPF S.A. (B1/B)

· Instrument: Senior unsecured notes, 144A / Reg S

· Size: US$1.2 billion

· Maturity: June 18, 2035, bullet, 8.75 years

· Coupon: 7.550%, semiannual, 30/360

· Yield: 7.85%, from IPT in the low 8% area

· Issue Price: $98.075

· Settlement: September 18, 2026

· Optional Redemption: Make-whole prior to December 18, 2030; callable at $103.775 from December 18, 2030, $101.8875 from December 18, 2031, and par from December 18, 2032

· Use of Proceeds: Fund the tender offers for the 6.950% 2027s and 2.500%/9.000% Step Up Amortizing 2029s, with remaining proceeds for general corporate purposes

· Bookrunners: BBVA, Itaú (B&D), J.P. Morgan, Santander. Joint bookrunner: Balanz

Read more

We maintain our Neutral recommendation on Simpar (Ba3/BB-/BB-) at the issuer level and move our preferred bond expression to the MOVIBZ 9.700% 2033 notes from the MOVIBZ 5.250% 2031s. This is a different Neutral from 1Q26. Then, improving fundamentals were offset by spreads that had already compressed substantially. Today, the bonds are materially wider and the operating evidence is stronger, but thin debt-service capacity remains a central constraint and has become more consequential as the expected decline in Brazilian rates has been pushed further out. Simpar still directs roughly half of EBITDA and more than four-fifths of EBIT to interest. We think that leaves too little financial headroom for an issuer-level Overweight, even as the repricing creates attractive bond-specific opportunities.

The operating improvement is real. Adjusted EBITDA increased 5.3% QoQ and 15.7% YoY to R$3.36 billion, while the EBITDA margin reached 29.9%. Service revenue strengthened as asset sales declined, improving the earnings mix. Gross leverage fell to 4.3x from 4.6x, net leverage to 3.2x from 3.5x and bond covenant leverage to 2.8x. Holding-company net debt also declined to R$1.4 billion and would have been approximately R$978 million pro forma for Porto Aratu, while R$350 million of the R$1.0 billion 2026 debt-repurchase target had been completed by July. These changes improve the downside profile and materially reduce the weight we assign to holdco refinancing risk.

Coverage, however, remains thin. EBITDA interest coverage improved to 1.98x from 1.86x in 1Q26, while EBIT interest coverage improved to 1.24x from 1.16x. For an asset-intensive leasing group, we place particular weight on EBIT coverage because depreciation and recurring fleet replacement represent real economic costs. At 1.24x, interest still absorbs more than four-fifths of EBIT, leaving limited room for weaker utilization, higher funding costs or another period of elevated investment. We do not think the modest improvement is enough to change the debt-service constraint in a higher-for-longer Brazilian rate environment.

Read more

We maintain our Neutral recommendation on Suzano (Baa3/BBB-/BBB-) at the issuer level and keep the bond-level Overweight on the 3.125% 2032 notes as our preferred expression of the credit. The issuer stance reflects a credit with a credible deleveraging path but balance-sheet repair still to deliver following the Arbex acquisition. The bond-level call remains a valuation decision. The 2032s offer the clearest relative value in the belly, combining additional spread and yield with a materially lower dollar price than the 2031s.

The case for Neutral at the issuer level sits in the gap between the leverage ratios and the cash-flow trend. Gross leverage rose to 5.0x and net leverage to 3.6x in 2Q26 as LTM EBITDA declined. Pro forma for Arbex, leverage should remain elevated, requiring further balance-sheet repair. Set against that, net debt declined during the quarter, free operating cash flow turned positive, liquidity covered short-term debt by roughly 6.9x at quarter-end and remains strong after allowing for the US$1.3 billion Arbex payment, and management has explicitly placed deleveraging ahead of further buybacks and discretionary cash investments. We see an improving deleveraging path, but not yet enough balance-sheet progress to support an issuer-level Overweight.

Read more

We maintain our Neutral recommendation on Telecom Argentina (B2/B/B-) and move to a bond-level Overweight on the TECOAR 8.500% 2036 notes, which remain our preferred expression within the curve. The issuer stance is a valuation decision rather than a cautious view on the direction of the credit. Telecom exits 2Q26 with stronger profitability, lower leverage, positive free cash flow and a maturity profile that allows us to assess liquidity as adequate rather than weak. The risk has shifted. Regulatory approval of the TMA acquisition is no longer the question. The question is how much of the operating improvement Telecom preserves after executing the required customer, spectrum and network divestitures. Against that backdrop, we think the credit continues to improve, but current valuations do not support a broad issuer-level Overweight.

The fundamental case is stronger than when we moved to Neutral. Profitability improved at both platforms in 2Q26, with broad cost discipline supporting earnings across the enlarged group. Net leverage declined to 1.48x and free cash flow remained positive despite elevated network investment. We also revise our liquidity assessment to adequate from weak. Headline short-term coverage remains below 1.0x, but Telecom faces only around US$330 million of principal maturities during the remainder of 2026 against US$645 million of cash and financial investments, close to 2.0x coverage of scheduled principal. Average debt life has lengthened to 4.8 years after another quarter of refinancing and prepayments. Liquidity remains the principal balance-sheet constraint, in our view, but it is no longer the central risk to the recommendation.

We would not treat the 2Q26 margin or free cash flow as a new run rate. Network investment remains elevated, and some normalization in cash generation would not change our view while leverage stays low and refinancing requirements remain manageable.

Read more

We maintain our Neutral recommendation on MercadoLibre (Baa3/BBB-/BBB-) and shift our preferred expression back to the MELI 3.125% 2031 notes from the 4.900% 2033s. The issuer stance is unchanged, but the reason for it has evolved. Our focus is no longer principally on the direction of margins. The more important question is whether growth is translating into enough earnings and cash to absorb the funding requirements of Fintech. 2Q26 showed operating profitability stabilizing, with adjusted EBITDA margin broadly unchanged sequentially at 9.6%, but the credit profile did not improve with it. Gross and net leverage rose to 3.39x and 1.65x, the credit portfolio grew 75.2% YoY against 49.8% consolidated revenue growth, and only $214 million of adjusted free cash flow remained after a quarter that generated $3.22 billion of free operating cash flow. We think that combination keeps MercadoLibre a strong investment-grade LatAm credit, but not one where current spreads offer enough cushion to move beyond Neutral.

For the issuer call, the question is whether current spreads compensate investors for a balance sheet expanding faster than earnings. We do not think they do. The 2031s at 128 bps OAS trade flat to the EM BBB corporate index and only 6 bps inside their own 1Y average, while sitting 5 bps wide of the 3M average of 123 bps. Both bonds also trade materially through the LatAm BBB curve. At those levels, the market already gives MercadoLibre substantial credit for its scale, ecosystem economics, and market access even as balance-sheet intensity continues to rise. Short-term coverage declined to 0.97x, or 1.08x including the undrawn revolver, and the credit portfolio continues to grow at a double-digit sequential pace. We see no issuer-level dislocation that would justify an Overweight.

Read more

We move to Neutral on JBS (Baa3/BBB-/BBB-) from Overweight but retain a bond-level Overweight on the JBS 7.250% 2053 notes, making them our preferred expression over the 6.500% 2052s. We also shift our standard below-par alternative to the 6.375% 2066s, while retaining a separate deep-discount expression for price-focused mandates. This is a valuation decision before it is a credit decision. Our 1Q26 Overweight rested on a curve that paid investors across maturities for a leverage cycle we expected to prove temporary. Much of that compensation has now been collected where it was largest. The 2031 to 2034 bonds trade 12 to 19 bps inside their own 1-year averages, while the 2052-and-longer bonds remain within 8 bps of their respective 1-year averages.

Weaker credit metrics raise the bar for an issuer-level Overweight, but they are not the reason we step back. The sequential recovery in 2Q26 and the return to positive reported free cash flow reduce the concern created by the prior quarter. The leverage cycle has simply lasted longer than we assumed when we made the call. Company-defined net leverage of 3.10x now sits outside management’s 2.0x to 3.0x target, and our lease-inclusive measure of 3.45x sits further above it. We remain comfortable owning the credit because the liability structure gives JBS time. Average debt maturity is above 15 years, near-term maturities remain limited through 2030, and the liquidity buffer was strengthened again in August. JBS can carry an extended period of weak earnings without the balance sheet forcing a decision.

Read more

We maintain our Neutral recommendation on Braskem and continue to prefer the BRKME 8.500% 2031s as the most balanced expression of risk across the senior unsecured curve. The basis for the call has changed since 1Q26. Braskem is no longer a stressed liquidity credit facing the possibility of a capital structure solution. Following the June creditor protection measures, the Chapter 15 filing, the expiration of applicable cure periods and subsequent payment defaults, the company is in an active restructuring process. What determines bond value from here is the form of the eventual solution, the treatment of existing unsecured claims, the amount and seniority of any new money and ultimate recovery. The move into default changes how we value the bonds, but it does not on its own justify a more negative stance when current dollar prices already discount substantial impairment.

The 2Q26 earnings recovery does not alter that framework. Adjusted EBITDA of R$5.25 billion was driven by a supply shock to petrochemical spreads that management itself does not treat as a cyclical turn, and reference spreads had already returned to pre-conflict levels by quarter-end. For creditors, the more important point is that the earnings did not translate into positive free operating cash flow. Working capital absorbed R$2.76 billion, reported operating cash flow was only R$29 million and free operating cash flow remained negative at R$535 million. The quarter shows what the assets can earn under favorable spreads. It does not show that the balance sheet problem has been solved. Available corporate cash covers only around two-thirds of remaining 2026 contractual maturities before the reclassification of defaulted obligations, which in our view makes some combination of creditor concessions and external support increasingly difficult to avoid.

Two developments since 1Q26 partially offset the fundamental weakness and are the main reason we do not move to Underweight. First, the existing capital structure still leaves the unsecured curve with recovery optionality. The 2Q26 debt breakdown shows corporate debt concentrated in bonds and unsecured debentures, so the senior unsecured curve has no large layer of collateralized debt ahead of it today. That may not last. Braskem’s August 20 filing confirms that creditor proposals under evaluation include potential capitalization and the granting of security interests over assets, which makes the amount and form of any new collateralization an important recovery variable.

Second, the Braskem Idesa agreement is a useful reference for the restructuring process, though not for recovery. It shows that Braskem can close a complex creditor negotiation involving substantial deleveraging, new capital and a prepackaged filing. Against that, the parent retains a funding commitment to the subsidiary that is an additional call on liquidity at a point when liquidity is the binding constraint, and we do not read the Idesa outcome as a recovery floor for BRKME unsecured notes. Braskem said on August 20 that discussions with financial creditors had intensified and that it was evaluating potential additional protection measures as the 60-day stay approaches expiry, with no decision yet reached on restructuring terms. The expiry of that protection period and the choice of restructuring path are the next decision points.

Read more

We upgrade Minerva (Ba3/BB/BB) to Overweight from Neutral and prefer the BEEFBZ 4.375% 2031 notes as the cleanest risk-adjusted expression of the credit. The upgrade is valuation-driven and does not rest on a clean fundamental inflection. Credit metrics weakened sequentially, working capital remains a drag on cash generation, and supplier finance has become a more meaningful component of adjusted leverage. Liquidity nevertheless remains strong, conventional net leverage is still materially below year-ago levels, and the intermediate bonds have repriced substantially more than the broader EM BB market. We think current spreads now provide sufficient compensation for the remaining credit risks.

The fundamental picture remains mixed. Net leverage increased to 2.94x from 2.77x, while cash interest coverage declined to 1.66x from 1.93x. Supplier finance increased R$744 million QoQ to R$4.70 billion, more than reversing the R$399 million decline recorded in 1Q26 and lifting adjusted net leverage to 3.90x from 3.56x. We therefore view interest coverage and adjusted leverage, rather than liquidity or refinancing risk, as the principal balance-sheet constraints. The increase in gross debt largely reflected deliberate prefunding, while net debt rose only R$667 million sequentially. Cash covers short-term debt by approximately 2.7x and is sufficient to cover scheduled maturities through 2029.

Cash conversion is the main fundamental swing factor. Free cash flow before working capital was positive in 2Q26, with the reported outflow driven by a R$1.07 billion working-capital investment. Management expects part of the first-half receivables and inventory build to begin converting into cash in 3Q26 and indicated that roughly R$3 billion of inventory could be monetized in the near term, potentially reducing leverage by 0.5x to 0.6x. We expect second-half cash generation to improve, but we do not underwrite management’s 2.2x to 2.3x year-end leverage path. China’s import quota is already a constraint on Brazilian export flexibility in 3Q26, and the risk, in our view, extends beyond the timing of the working-capital release. A tighter Brazilian cattle cycle could keep earnings and cash conversion constrained into 2027 even if inventories normalize in 2H26, limiting the pace of subsequent deleveraging. Strong global beef pricing and Minerva’s export platform should provide some support, but we do not assume they fully offset higher cattle costs.

Read more

We maintain our Opportunistic Overweight recommendation on CSN and shift our preferred bond to the CSN Resources 8.875% 2030s. We no longer recommend the CSN Resources 4.625% 2031s as our preferred expression, although we continue to view them as appropriate for accounts placing greater weight on downside protection than carry. We do not recommend the CSN Resources 5.875% 2032s or the residual CSN Inova 6.750% 2028s. The Opportunistic qualifier reflects CSN’s weak organic cash generation, tight parent-level liquidity and continued dependence on liability management and asset monetization, but we think current spreads overcompensate investors for the risk of principal impairment over our 9 to 12 month horizon.

CSN’s USD senior unsecured curve carries a CCC+ composite rating, with Moody’s at Caa1, S&P at B and Fitch at CCC+. Against selected CCC-category comparables, the spread differential remains unusually wide. The CSN Resources 8.875% 2030s at 1,349 bps OAS trade roughly 570 to 940 bps wide of the selected Kosmos, MHP and Metinvest bonds shown in our curve analysis. The differential has widened through 2026 rather than representing a stable cohort premium. The selected Kosmos and Metinvest bonds are up approximately 27% to 62% YTD, while the CSN bonds are down approximately 7.5% to 14.9%, reinforcing that the spread gap is primarily issuer-specific rather than a generic lower-rated credit technical. Metinvest provides the clearest illustration. Despite direct exposure to the war in Ukraine, its 2027 and 2029 bonds trade roughly 720 to 780 bps inside the CSN Resources 2030s. We do not treat those spreads as fair-value targets, but the differential supports our view that CSN continues to price materially more credit stress than our base case assumes. The EM B index at 407 bps is more useful as a potential re-rating destination, leaving approximately 940 bps of spread between the CSN Resources 2030s and that benchmark.

Read more

We move to Neutral on YPF (B1/B/B-) from Overweight and shift our preferred expression to the YPFDAR 9.500% 2031 notes from the 8.250% 2034s. The issuer change is a valuation call rather than a credit call. YPF’s fundamental profile strengthened materially in 2Q26, with record EBITDA, net leverage below 1.2x, cash interest coverage above 6x and a liquidity position we now assess as adequate rather than weak. Our issue is not with the direction of the credit. In fact, we think the underlying credit is stronger than the published rating category alone suggests. Neutral should therefore not be read as a cautious view on YPF’s credit trajectory. The question is whether enough issuer-level upside remains after the re-rating already delivered.

The market is increasingly pricing YPF on its forward credit profile rather than its headline ratings. The 7.000% 2033s trade at 253 bps OAS, only 13 bps wide of the EM BB corporate index at 240 bps and well inside the EM B index at 402 bps. We think that is broadly justified. YPF’s stronger leverage, liquidity, cash generation and asset quality support a market-implied credit profile above the published single-B ratings, and we would not expect the bonds to revert toward EM B spreads.

However, the speed of the repricing matters. Since our 1Q26 report, the 2033s have tightened 54 bps, the 8.250% 2034s 53 bps, the 9.500% 2031s 92 bps and the 7.000% 2047s 48 bps, compared with only 8 bps for the EM BB index. The four lines we consider most relevant also trade inside their 3-month averages and just 13 to 29 bps above their 1-year lows, against 219 to 314 bps below their 1-year highs and 139 to 193 bps inside their 3-year averages. For an issuer-level Overweight from here, we would want either broad value across the curve or a fundamental trajectory that we think the market is still materially underestimating. At current levels, most of the curve already prices a substantial part of the improvement we expect, leaving the opportunity increasingly concentrated in bond selection.

Read more

We remain Overweight on PEMEX, but with lower conviction and a narrower set of preferred positions supporting the call. The 2Q26 case for owning the credit now rests on carry, a refinancing profile that has improved more than any other part of the credit, and the specific points on the curve where compensation remains adequate. It no longer rests on further PEMEX-specific compression versus the sovereign, which we think has largely played out.

The fundamental improvement is what keeps us at Overweight rather than Neutral. Gross leverage declined to 4.0x from 5.2x in 1Q26 and 6.8x a year earlier, net leverage fell to 3.6x from 4.7x, and interest coverage improved to 2.6x from 2.1x. Short-term debt declined 55.9% YoY to $12.70 billion, materially reducing the maturity wall that framed the credit through most of 2025. LTM FFO returned to positive territory at $589 million from a deficit of approximately $4.0 billion in 1Q26, reflecting both the stronger quarter and the roll-off of a materially negative 2Q25, but remained weak relative to PEMEX’s approximately $79.6 billion debt burden.

We would nonetheless resist reading the improvement as standalone deleveraging. Supplier obligations increased sequentially to $21.41 billion, while only 37.2% of the 2026 investment plan had been exercised by June, leaving a heavier spending burden in 2H26. Pre-contribution net free cash flow also remained negative at $635 million despite the sharp improvement in EBITDA. Management indicated that federal support for the remainder of 2026 is already established in the budget, with no additional amounts anticipated this year. We think that is the most relevant forward statement of the quarter for bondholders because it confirms the support while limiting expectations for incremental assistance during the remainder of the year.

Liquidity is where bondholder risk has changed most, even though we continue to characterize it as weak. Cash and short-term investments of $8.27 billion, together with roughly $4.44 billion of available committed facilities, cover short-term debt at approximately 1.0x, which remains thin against a $21.41 billion supplier balance and continued reliance on external funding. In our view, the renewal of the revolving facilities matters more than the coverage ratio itself because it removes near-term bank refinancing as another source of uncertainty and extends committed funding visibility through 2029.

Read more

We maintain our Neutral recommendation on Gerdau and view the GGBRBZ 5.750% 2035 notes as the preferred expression of the credit. 2Q26 leaves us with an issuer that improved on nearly every metric we track and a curve that has already absorbed much of that improvement. Gross leverage of 1.32x, net leverage of 0.85x, FFO to gross debt of 57.6% and R$9.94 billion of liquidity against R$1.11 billion of short-term debt place Gerdau at the stronger end of the BBB category, while 2026 capex is running slightly below the R$4.7 billion plan. We stay Neutral because the constraint is price, not credit. At 138 bps OAS, the 2035s plot essentially on the duration-adjusted U.S. BBB curve, leaving little if any incremental spread compensation for Brazil risk.

We think that positioning is defensible but leaves little in reserve. North America generated roughly 74% of LTM adjusted EBITDA against 53% of gross debt denominated in U.S. dollars, providing a meaningful natural hedge and an earnings base that is substantially less exposed to Brazilian steel pricing. The 2035s trade approximately 54 bps inside the Brazil 2033 sovereign with about a year more duration, and we do not treat the sovereign as a valuation floor for an issuer with this earnings mix. The practical consequence for bondholders, in our view, is that near-term earnings sensitivity is now more heavily tied to U.S. steel spreads, Section 232 and the U.S. investment cycle, while Brazil remains the main source of potential incremental recovery rather than the current earnings anchor.

Read more

We remain Neutral on Cemex and maintain our Buy recommendation on the CEMEX 5.750% 2036 notes, which have been our preferred curve expression since issuance in June. As discussed in our new-issue report, Cemex Prices US$1.5bn 5.750% 2036 Senior Unsecured Notes, the original call reflected the bonds’ carry, liquidity, near-par entry point and more efficient access to the long end of the Cemex curve. Since then, secondary-market widening, stronger free cash flow and further deleveraging have increased our conviction in the Buy. The issuer remains Neutral because value is concentrated in the 2036s rather than broad enough across the curve to support an issuer-level Overweight.

Fundamentals continue to improve. Net leverage including perpetual notes declined to 2.1x from 2.3x, while the ex-perpetual measure increased modestly to 1.8x from 1.7x following the cash redemption of the subordinated notes. EBITDA interest coverage improved to 7.5x and FFO to gross debt reached 34.4%. The new $3.0 billion revolving facility, an average debt life of 5.4 years and approximately $741 million of maturities through 2028 compare with $2.95 billion of liquidity, making near-term refinancing risk a secondary consideration. We think Cemex is approaching metrics capable of supporting a one-notch upgrade, and Fitch revised the Outlook on its BBB- rating to Positive in April 2026, but our recommendation remains anchored in fundamentals and valuation.

Read more